Intelligent power monitoring management method and system for electrical cabinet

Abstract

The invention relates to the technical field of power monitoring, in particular to an intelligent power monitoring management method and system for an electrical cabinet, and the method comprises the steps: obtaining a state data point at each moment in the operation process of the electrical cabinet, the state data point comprising temperature data and residual current data; and detecting the state data points by using an isolated forest algorithm to obtain an abnormal score, and correcting the abnormal score to obtain an optimal abnormal score for monitoring the operation state of the electrical cabinet. According to the method, the initial abnormal score is corrected, so that the accuracy of the monitoring result of the isolated forest is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power monitoring, and in particular to an intelligent power monitoring and management method and system for electrical cabinets. Background Art

[0002] An electrical cabinet is a cabinet used to store and protect electrical equipment. It is usually made of metal or non-metallic materials, and various electrical components are installed inside, such as switches, relays, contactors, circuit breakers, transformers, control modules, etc. Its main function is to achieve centralized management, protection and control of electrical equipment. Monitoring and managing the electrical cabinet can timely detect and identify possible problems in the electrical cabinet, and timely anomaly detection helps to quickly respond at the beginning of the problem, thereby minimizing the damage caused by potential faults to the equipment to the greatest extent.

[0003] The Chinese patent application document with the publication number CN119010364A discloses an Internet of Things-based weak current control and monitoring system and method. This method collects monitoring data through monitoring nodes deployed on weak current equipment, and generates a tracking certificate for the monitoring data of the same weak current equipment at the same moment. The cloud analysis platform verifies the legitimate source of the received data through the tracking certificate to ensure that the data is not lost during transmission, and guarantees the authenticity and integrity of the monitoring data during transmission. Then, the system constructs an isolation forest with the help of the weak current equipment intelligent monitoring model, randomly selects temperature, current, and voltage features for splitting, continuously divides the data points until each data point is isolated, automatically outputs the anomaly points, and then the system identifies the potential faults of the equipment and intelligently controls and adjusts the working state of the weak current equipment to ensure the normal operation of the weak current equipment.

[0004] When performing power monitoring and management on an electrical cabinet, the isolation forest algorithm can be used to monitor the parameters of the electrical cabinet to understand its operating state. In the traditional isolation forest algorithm, the weights of each isolation tree are the same, but in fact, the detection effects of different isolation trees on abnormal data are different. Therefore, the accuracy of the abnormal scores obtained by the traditional isolation forest algorithm is relatively low, and it is impossible to effectively monitor and manage the electrical cabinet. Summary of the Invention

[0005] In order to solve the problem that the same weight of each isolation tree in the traditional isolation forest algorithm leads to relatively low accuracy of the monitoring results, the present invention provides an intelligent power monitoring and management method and system for electrical cabinets.

[0006] In the first aspect, the present invention provides an intelligent power monitoring and management method for electrical cabinets, adopting the following technical solutions: Obtain the status data points at each moment during the operation of the electrical cabinet. The status data points include temperature data and residual current data; use the Isolation Forest algorithm to detect the status data points to obtain anomaly scores, and correct the anomaly scores to obtain the optimal anomaly scores for monitoring the operation status of the electrical cabinet; Among them, the method for correcting the anomaly scores to obtain the optimal anomaly scores is as follows: Calculate the electrical environment anomaly index at each moment. The electrical environment anomaly index characterizes the correlation between the temperature data and the residual current data; calculate the anomaly index of each leaf node in each layer of the isolation tree. The anomaly index of the leaf node is positively correlated with the mean value of the electrical environment anomaly index corresponding to the moment within the leaf node; perform a weighted sum of the anomaly indices of each layer of leaf nodes to obtain the isolation anomaly factor of the isolation tree; normalize the isolation anomaly factor, and use the normalized isolation anomaly factor to perform a weighted sum of the initial anomaly scores to obtain the optimal anomaly scores.

[0007] Analyze the anomaly degree of the leaf nodes in the isolation tree based on the electrical environment anomaly index, and construct an isolation anomaly factor to evaluate the isolation effect of the isolation tree on the anomaly data; improve the calculation method of the anomaly scores in the Isolation Forest based on the isolation anomaly factor, so that the isolation tree with a good isolation effect on the anomaly data has a larger isolation anomaly factor, making the finally calculated anomaly scores more accurately reflect the abnormal state of the electrical cabinet and improving the accuracy of the monitoring results of the Isolation Forest.

[0008] Preferably, the method further includes: using multiple status data points to construct the near-neighbor operation status sequence corresponding to the moment. The near-neighbor operation status sequence includes a near-neighbor temperature data sequence and a near-neighbor residual current data sequence, input the near-neighbor operation status sequence into the Wilcoxon signed-rank test algorithm, and the hypothesis test condition is that there is an upward trend in the near-neighbor operation status sequence to obtain the p-value of the existence of an upward trend in the near-neighbor operation status sequence.

[0009] The p-value reflects the reliability of the existence of an upward trend in the near-neighbor residual current data sequence and the near-neighbor temperature data sequence.

[0010] Preferably, the method further includes: performing a first-order difference processing on the near-neighbor operation status sequence to obtain a difference sequence, and calculating the local upward intensity index of the near-neighbor operation status sequence at each moment. Among them, the expression is:

[0011] In the formula, represents the local upward intensity index of the near-neighbor operation status sequence at time t, represents the p-value of the near-neighbor operation status sequence at time t after passing the Wilcoxon signed-rank test, and respectively represent the number of positive and negative elements in the difference sequence corresponding to the neighboring operation state sequence at time t, represents the a-th positive value in the difference sequence corresponding to the neighboring operation state sequence at time t, represents the b-th negative value in the difference sequence corresponding to the neighboring operation state sequence at time t.

[0012] The change trend of the neighboring operation state sequence can be reflected by the local rising intensity index, providing a theoretical basis for analyzing the electrical environment anomaly index.

[0013] Preferably, the method further includes: obtaining the maximum value points of the neighboring temperature data sequence and the neighboring residual current data sequence, constructing the neighboring residual current maximum value sequence and the neighboring temperature maximum value sequence, and inputting the neighboring residual current maximum value sequence and the neighboring temperature maximum value sequence into the DTW algorithm to obtain two maximum value sequences.

[0014] Preferably, the expression of the electrical environment anomaly index is:

[0015] In the formula, represents the electrical environment anomaly index of the state data point at time t, and respectively represent the local rising intensity indexes of the neighboring residual current sequence and the neighboring temperature sequence at time t, represents the length of the maximum value sequence after aligning the neighboring residual current maximum value sequence and the neighboring temperature maximum value sequence at time t using the DTW algorithm, represents the Manhattan distance of the c-th maximum value point in the two aligned sequences.

[0016] The difference between the neighboring residual current sequence and the neighboring temperature sequence can be reflected by the electrical environment anomaly index, further reflecting the abnormal degree of the electrical environment.

[0017] Preferably, the expression of the anomaly index of the leaf node is:

[0018] In the formula, represents the anomaly index of the leaf node in the h-th layer of the isolation tree, represents the number of state data points in the leaf node in the h-th layer of the isolation tree, represents the electrical environment anomaly index of the m-th state data point in the h-th layer of the isolation tree, represents the maximum value of the electrical environment anomaly indexes of all state data points in the isolation tree.

[0019] The abnormal index of the leaf node is calculated through the status data points, which improves the accuracy of the calculation result and provides a theoretical basis for calculating the isolated abnormal factor.

[0020] Preferably, the expression of the isolated abnormal factor of the isolation tree is:

[0021] In the formula, represents the isolated abnormal factor of the i-th isolation tree in the isolation forest, represents the maximum depth of the i-th isolation tree, h represents the leaf node of the h-th layer in the isolation tree, represents the abnormal index of the electrical environment of the leaf node of the h-th layer of the i-th isolation tree.

[0022] Preferably, the expression of the best abnormal score is:

[0023] In the formula, represents the best abnormal score of the operating state of the electrical cabinet at time v, N represents the number of isolation trees set in the isolation forest algorithm, represents the isolated abnormal factor of the i-th isolation tree in the isolation forest algorithm, represents the maximum value of the isolated abnormal factors of all isolation trees in the isolation forest algorithm, represents the initial abnormal score of the operating state data of the electrical cabinet at time v in the i-th isolation tree.

[0024] The best abnormal score is obtained by correcting the initial abnormal score. Compared with the fixed value weight of the traditional isolation tree, the accuracy of the abnormal score is improved, and the operating state of the electrical cabinet can be accurately reflected.

[0025] Preferably, the method further includes the step of filtering and denoising the status data points.

[0026] In a second aspect, the present invention provides an intelligent power monitoring and management system for an electrical cabinet, adopting the following technical solutions: An intelligent power monitoring and management system for an electrical cabinet includes: a processor and a memory, and the memory stores computer program instructions, which implement an intelligent power monitoring and management method for an electrical cabinet according to the above when the computer program instructions are executed by the processor.

[0027] Generate a computer program for the above intelligent power monitoring and management method for an electrical cabinet and store it in the memory to be loaded and executed by the processor, so as to make a system according to the memory and the processor for convenient use.

[0028] The present invention has the following technical effects: Construct a local rising intensity index by analyzing the rising trend intensity of the residual current data and temperature data of the circuit during the operation of the electrical cabinet, and respectively evaluate the abnormal influence degree of the residual current and temperature in the circuit on the operation state of the electrical cabinet; Since the magnitude of the residual current in the circuit will cause corresponding changes in temperature, analyze the difference degree between the local rising intensity indexes of the residual current and temperature and the similarity degree of the peak value changes of the residual current and temperature within the local time to construct an electrical environment anomaly index, and evaluate the effectiveness of the electrical cabinet cooling system; Analyze the anomaly degree of the leaf nodes in the isolation tree based on the electrical environment anomaly index, and construct an isolation anomaly factor to evaluate the isolation effect of the isolation tree on the abnormal data; Improve the calculation method of the abnormal score in the isolation forest based on the isolation anomaly factor, so that the isolation tree with a good isolation effect on the abnormal data has a larger isolation anomaly factor, and the finally calculated abnormal score can more accurately reflect the abnormal state of the electrical cabinet, and take different measures according to different abnormal degrees to realize the intelligent monitoring and management of the electrical power of the electrical cabinet. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a flowchart of an intelligent electrical power monitoring and management method for an electrical cabinet according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] An embodiment of the present invention discloses an intelligent electrical power monitoring and management method for an electrical cabinet. Refer to Figure 1 , and includes the following steps, specifically as follows: S1: Obtain the state data points at each moment during the operation of the electrical cabinet.

[0032] The status data points include temperature data and residual current data. During the electrical operation, the temperature data of one point on the surface of the electrical cabinet and the residual current data of the electrical cabinet cable are collected through a temperature sensor and a current transformer respectively. The collection interval is set to 1 second. The mean filter denoising algorithm is used to denoise the collected temperature data and residual current data, and the standard deviation normalization method is used to normalize the denoised data. The mean filter denoising algorithm and the standard deviation normalization method are well-known technologies and will not be elaborated here. The residual current data is the leakage current. In the normal working circuit of the electrical cabinet, the vector sum of the currents between the phase line and the neutral line is 0. If the vector sum of the currents is not 0, it indicates that there is a leakage phenomenon in the circuit and a residual current is generated. In other embodiments, the status data points can also be power data and temperature data.

[0033] S2: Calculate the local upward intensity index of the neighboring operating state sequence at each moment.

[0034] Use the status data points of the first n moments before the t-th moment to construct the neighboring operating state sequence at the corresponding moment. The neighboring operating state sequence includes the neighboring temperature data sequence and the neighboring residual current data sequence. For example, use the temperature data of the first 30 moments before the t-th moment to construct the neighboring temperature data sequence, and use the residual current data of the first 30 moments before the t-th moment to construct the neighboring residual current data sequence.

[0035] Input the neighboring operating state sequence into the Wilcoxon signed-rank test algorithm. The hypothesis test condition is that there is an upward trend in this neighboring operating state sequence, and obtain the p-value indicating that there is an upward trend in this neighboring operating state sequence. It can be understood that inputting the neighboring temperature data sequence into the Wilcoxon signed-rank test algorithm obtains a p-value, and inputting the neighboring residual current data sequence into the Wilcoxon signed-rank test algorithm obtains a p-value.

[0036] Perform a first-order difference processing on the neighboring operating state sequence to obtain a difference sequence, and calculate the local upward intensity index of the neighboring operating state sequence at each moment. Among them, the expression is:

[0037] In the formula, represents the local upward intensity index of the neighboring operating state sequence at the t-th moment, represents the p-value after the neighboring operating state sequence at the t-th moment passes the Wilcoxon signed-rank test, and respectively represent the number of positive and negative elements in the difference sequence corresponding to the neighboring operating state sequence at the t-th moment, represents the a-th positive value in the difference sequence corresponding to the neighboring operating state sequence at the t-th moment, It represents the b-th negative value in the difference sequence corresponding to the near-neighbor operating state sequence at time t. The 1 in the denominator represents a hyperparameter to avoid the phenomenon of a zero denominator.

[0038] It can be understood that the difference sequence includes the temperature data difference sequence and the residual current data difference sequence. The near-neighbor temperature data sequence corresponds to a local rising intensity index, and the near-neighbor residual current data sequence corresponds to a local rising intensity index.

[0039] During the operation of the electrical cabinet, the magnitude of the residual current indicates the severity of the leakage. Generally, all electrical systems have residual current. The trace residual current caused by non-electrical faults has little harm to the electrical system and can be considered a normal phenomenon. However, as the residual current increases, it will cause the temperature of the line to rise. When the temperature reaches a certain level, it may cause the cable to burn, resulting in electrical-related accidents.

[0040] Based on the above analysis, by constructing the local rising intensity index to respectively reflect the trend intensity of the residual current data and the temperature data. In the near-neighbor residual current sequence at time t, if there is an upward trend in the residual current in the electrical cabinet, that is is relatively large, and when the degree of increase is greater than the degree of decrease, that is is relatively large, it indicates that there is an abnormal power operating state in the electrical cabinet at this time, so the calculated local rising intensity index is relatively large.

[0041] S3: Calculate the electrical environment anomaly index at each moment. The electrical environment anomaly index characterizes the correlation between the temperature data and the residual current data.

[0042] When residual current is generated in the electrical cabinet, according to Joule's law, it can be obtained that there will be a certain thermal effect in the circuit, and the greater the residual current, the more heat is generated, which in turn causes the temperature to rise. Therefore, at the moment when the residual current is relatively large, the corresponding temperature is relatively high. If the temperature in the electrical cabinet rises and is not controlled, a flashover phenomenon may occur in a short time, causing great losses to the electrical cabinet and surrounding environmental facilities. Therefore, based on the local rising intensity index, an electrical environment anomaly index is constructed to reflect the degree of abnormality in the environment during the operation of the electrical cabinet.

[0043] Obtain the maximum value points of the near-neighbor temperature data sequence and the near-neighbor residual current data sequence, and construct the near-neighbor residual current maximum value sequence and the near-neighbor temperature maximum value sequence. Input the near-neighbor residual current maximum value sequence and the near-neighbor temperature maximum value sequence into the DTW algorithm to obtain two maximum value sequences.

[0044] The expression of the electrical environment anomaly index is:

[0045] In the formula, The electrical environment anomaly index representing the state data point at time t and respectively represent the local rising intensity indexes of the near - neighbor residual current sequence and the near - neighbor temperature sequence at time t represents the length of the maximum - value sequence after aligning the near - neighbor residual current maximum - value sequence and the near - neighbor temperature maximum - value sequence at time t using the DTW algorithm (Dynamic Time Warping) represents the Manhattan distance of the c - th maximum - value point in the two sequences after alignment. The Manhattan distance can also be understood as the absolute value of the difference between the two maximum - value points

[0046] If the difference in the local rising intensity indexes of the near - neighbor residual current sequence and the near - neighbor temperature sequence at time t is smaller, that is the smaller it is, it indicates that the residual current data and the temperature data have similar trend intensities. At the same time, if the Manhattan distance of the two maximum - value points at the same position in the near - neighbor residual current maximum - value sequence and the near - neighbor temperature maximum - value sequence after alignment is smaller, that is the smaller it is, it indicates that the change in the residual current can cause a corresponding change in the temperature. That is, at this time, the cooling system of the electrical cabinet is poor, unable to control the temperature well, the operating environment of the electrical cabinet is poor, and relevant power accidents are likely to occur. Therefore, the calculated electrical environment anomaly index is larger

[0047] S4: Calculate the anomaly index of each leaf node in each layer of the isolation tree. The anomaly index of the leaf node is positively correlated with the mean value of the electrical environment anomaly index corresponding to the time in the leaf node

[0048] Construct an isolation forest model, where the number of isolation trees is 100 and the depth of the isolation tree is 10 The expression for the anomaly index of the leaf node is

[0049] In the formula represents the anomaly index of the leaf node in the h - th layer of the isolation tree represents the number of state data points in the leaf node in the h - th layer of the isolation tree represents the electrical environment anomaly index of the m - th state data point in the h - th layer of the isolation tree represents the maximum value of the electrical environment anomaly indexes of all state data points in the isolation tree reflects the relative abnormal operation degree of the electrical cabinet at the m - th moment in the isolation tree

[0050] S5: Weight - sum the anomaly indexes of each layer of leaf nodes to obtain the isolation anomaly factor of the isolation tree

[0051] The expression for the isolation anomaly factor of the isolation tree is

[0052] In the formula, represents the isolation anomaly factor of the i-th isolated tree in the isolation forest, represents the maximum depth of the i-th isolated tree, h represents the leaf node of the h-th layer in the isolated tree, represents the electrical environment anomaly index of the leaf node of the h-th layer of the i-th isolated tree.

[0053] In the isolation forest algorithm, if the isolated tree has a good recognition effect on abnormal data, the data with a greater degree of abnormality will be recognized earlier. That is, when h is smaller, the degree of abnormality of the leaf node of the h-th layer should be greater. Therefore, a larger weight is given to the electrical environment anomaly index of the leaf node with a smaller depth of the isolated tree.

[0054] Therefore, when h is smaller and the degree of abnormality of the leaf node is greater, it indicates that the isolated tree has a good isolation effect on abnormal data. When this isolated tree is used to identify abnormal data, a larger weight should be given, and thus the calculated isolation anomaly factor is larger.

[0055] S6: Use the isolation forest algorithm to detect the state data points to obtain the abnormal score, and correct the abnormal score to obtain the optimal abnormal score for monitoring the operating state of the electrical cabinet.

[0056] Normalize the isolation anomaly factor, and use the normalized isolation anomaly factor to weighted sum the initial abnormal scores to obtain the optimal abnormal score.

[0057] The expression of the optimal abnormal score is:

[0058] In the formula, represents the optimal abnormal score of the operating state of the electrical cabinet at time v, N represents the number of isolated trees set in the isolation forest algorithm, represents the isolation anomaly factor of the i-th isolated tree in the isolation forest algorithm, represents the maximum value of the isolation anomaly factors of all isolated trees in the isolation forest algorithm, is used for normalizing the isolation anomaly factor, represents the initial abnormal score of the operating state data of the electrical cabinet at time v in the i-th isolated tree.

[0059] If the degree of abnormality of the electrical cabinet at time v is higher, it is more easily isolated in the isolated tree, and the corresponding abnormal score is larger. Therefore, the calculated optimal abnormal score is higher.

[0060] If the abnormal score of the operating state of the electrical cabinet at time v obtained after improving the isolation forest algorithm is greater than or equal to the first-level abnormal threshold, it indicates that the operating state of the electrical cabinet is severely abnormal at this time, and a first-level early warning should be issued in a timely manner to prompt the relevant staff to handle it immediately to avoid serious electrical accidents; if the obtained abnormal score is less than the first-level abnormal threshold and greater than or equal to the second-level abnormal threshold, it indicates that the operating state of the electrical cabinet is relatively abnormal at this time, and a second-level early warning should be issued to prompt the relevant staff to handle it in a timely manner; if the obtained abnormal score is less than or equal to the second-level abnormal threshold, it indicates that the operating state of the electrical cabinet is slightly abnormal or normal. In this way, the intelligent monitoring and management of the power of the electrical cabinet is realized. Among them, the first-level abnormal threshold and the second-level abnormal threshold are selected according to the actual situation. Exemplarily, the first-level abnormal threshold is set to 0.8, and the second-level abnormal threshold is set to 0.5.

[0061] The embodiment of the present invention also discloses an intelligent power monitoring and management system for an electrical cabinet, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent power monitoring and management method for an electrical cabinet according to the present invention is implemented.

[0062] The above system also includes other components well-known to those skilled in the art, such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be described in detail here.

[0063] The above are all the preferred embodiments of the present invention. The protection scope of the present invention is not limited by this. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. An intelligent power monitoring and management method for an electrical cabinet, characterized in that, Including the steps: Obtain the state data points at each moment during the operation of the electrical cabinet. The state data points include temperature data and residual current data; use the Isolation Forest algorithm to detect the state data points to obtain an anomaly score, and correct the anomaly score to obtain the optimal anomaly score for monitoring the operation state of the electrical cabinet; Among them, the method for correcting the anomaly score to obtain the optimal anomaly score is: Calculate the electrical environment anomaly index at each moment. The electrical environment anomaly index characterizes the correlation between the temperature data and the residual current data; calculate the anomaly index of each leaf node in each layer of the isolation tree. The anomaly index of the leaf node is positively correlated with the mean value of the electrical environment anomaly index corresponding to the corresponding moment in the leaf node; perform a weighted sum of the anomaly indices of each layer of leaf nodes to obtain the isolation anomaly factor of the isolation tree; normalize the isolation anomaly factor, and use the normalized isolation anomaly factor to perform a weighted sum of the initial anomaly scores to obtain the optimal anomaly score.

2. An intelligent power monitoring and management method for an electrical cabinet according to claim 1, characterized in that, The method further includes: constructing a neighboring operating state sequence corresponding to the corresponding moment using multiple state data points. The neighboring operating state sequence includes a neighboring temperature data sequence and a neighboring residual current data sequence, and inputting the neighboring operating state sequence into the Wilcoxon signed-rank test algorithm. The hypothesis testing condition is that there is an upward trend in the neighboring operating state sequence, and obtain the p-value that there is an upward trend in the neighboring operating state sequence.

3. The intelligent power monitoring and management method for an electrical cabinet according to claim 2, wherein, The method further includes: performing a first-order difference processing on the neighboring operating state sequence to obtain a difference sequence, and calculating the local upward intensity index of the neighboring operating state sequence at each moment, where the expression is: In the formula, represents the local upward intensity index of the neighboring operation state sequence at time t, represents the p-value after the Wilcoxon signed-rank test of the neighboring operation state sequence at time t, and respectively represent the number of positive and negative elements in the difference sequence corresponding to the neighboring operation state sequence at time t, represents the a-th positive value in the difference sequence corresponding to the neighboring operation state sequence at time t, represents the b-th negative value in the difference sequence corresponding to the neighboring operation state sequence at time t.

4. An intelligent power monitoring and management method for an electrical cabinet according to claim 3, characterized in that, The method further includes: obtaining the maximum value points of the neighboring temperature data sequence and the neighboring residual current data sequence, and constructing a neighboring residual current maximum value sequence and a neighboring temperature maximum value sequence, and inputting the neighboring residual current maximum value sequence and the neighboring temperature maximum value sequence into the DTW algorithm to obtain two maximum value sequences.

5. An intelligent power monitoring and management method for an electrical cabinet according to claim 4, characterized in that, The expression of the electrical environment anomaly index is: In the formula, represents the electrical environment anomaly index of the state data point at time t, and respectively represent the local rising intensity indices of the near-neighbor residual current sequence and the near-neighbor temperature sequence at time t, represents the length of the maximum value sequence after aligning the near-neighbor residual current maximum value sequence and the near-neighbor temperature maximum value sequence at time t using the DTW algorithm, represents the Manhattan distance of the c-th maximum value point in the two aligned sequences.

6. An intelligent power monitoring and management method for an electrical cabinet according to claim 1, characterized in that, The expression of the anomaly index of the leaf node is: In the formula, represents the anomaly index of the leaf node in the h-th layer of the isolation tree, represents the number of status data points in the leaf node in the h-th layer of the isolation tree, represents the electrical environment anomaly index of the m-th status data point in the h-th layer of the isolation tree, represents the maximum value of the electrical environment anomaly indices of all status data points in the isolation tree.

7. An intelligent power monitoring and management method for an electrical cabinet according to claim 1, characterized in that, The expression of the isolation anomaly factor of the isolation tree is: In the formula, represents the isolation anomaly factor of the i-th isolation tree in the isolation forest, represents the maximum depth of the i-th isolation tree, h represents the leaf node of the h-th layer in the isolation tree, represents the electrical environment anomaly index of the leaf node of the h-th layer of the i-th isolation tree.

8. An intelligent power monitoring and management method for an electrical cabinet according to claim 7, characterized in that, The expression of the optimal anomaly score is: Wherein, represents the optimal anomaly score of the operating state of the electrical cabinet at time v, N represents the number of isolation trees set in the isolation forest algorithm, represents the isolation anomaly factor of the i-th isolation tree in the isolation forest algorithm, represents the maximum value of the isolation anomaly factors of all isolation trees in the isolation forest algorithm, represents the initial anomaly score of the operating state data of the electrical cabinet at time v in the i-th isolation tree.

9. The intelligent power monitoring and management method for an electrical cabinet according to claim 1, characterized in that The method further includes the step of filtering and denoising the state data points.

10. An intelligent power monitoring and management system for an electrical cabinet, characterized in that, Including: A processor and a memory, where the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent power monitoring and management method for an electrical cabinet according to any one of claims 1-9 is implemented.

Images